Tilted-Mode All-Optical Diffractive Deep Neural Networks
Diffractive deep neural networks (D<sup>2</sup>NNs) typically adopt a densely cascaded arrangement of diffractive masks, leading to multiple reflections of diffracted light between adjacent masks, thereby affecting the network’s inference capability. It is challenging to fully simulate t...
Saved in:
| Main Authors: | , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
MDPI AG
2024-12-01
|
| Series: | Micromachines |
| Subjects: | |
| Online Access: | https://www.mdpi.com/2072-666X/16/1/8 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| Summary: | Diffractive deep neural networks (D<sup>2</sup>NNs) typically adopt a densely cascaded arrangement of diffractive masks, leading to multiple reflections of diffracted light between adjacent masks, thereby affecting the network’s inference capability. It is challenging to fully simulate this multiple-reflection phenomenon. To eliminate this phenomenon, we designed tilted-mode all-optical diffractive deep neural networks (T-D<sup>2</sup>NNs) and proposed a theoretical model for diffraction propagation in the tilted mode. Simulation results indicate that T-D<sup>2</sup>NNs address the performance degradation caused by interlayer reflections in D<sup>2</sup>NNs constructed with high-index diffractive masks. In classification tasks, T-D<sup>2</sup>NNs achieve better classification results compared to D<sup>2</sup>NNs that consider interlayer reflections. |
|---|---|
| ISSN: | 2072-666X |